Connect an MCP client
The proxium MCP server is at https://proxium.tech/mcp. It gives an AI agent the memory of your proxium projects. The agent reads what the team and earlier sessions learned, and it saves what a later session needs.
The server uses Streamable HTTP. It signs you in with OAuth, and the agent then acts as you. It can use the projects that you are a member of. It does not use a virtual key.
Turn memory on
Do this one time for each project. Each member of the project can do it.
- In the console, open Settings.
- In the Memory section, set Memory to on.
While memory is off, proxium refuses to save or correct a memory. The tool gets the message memory is off for this project. Turn it on under Settings.
Add the server to your client
The URL is the same for each client. Only the place where it goes is different.
Claude Code
Run this command. Then run /mcp in a session to sign in.
claude mcp add --transport http proxium https://proxium.tech/mcp
Claude Desktop and claude.ai
- Open Customize > Connectors.
- Select Add custom connector.
- Paste
https://proxium.tech/mcp, and select Add.
On a Team or Enterprise plan, an owner adds the connector under Organization settings > Connectors. Each member then selects Connect.
Codex
To start each task with the memory, add the skills or the project rule as well.
codex mcp add proxium --url https://proxium.tech/mcp
codex mcp login proxium
Cursor
Add the server to .cursor/mcp.json for one project, or to ~/.cursor/mcp.json for all projects. Cursor asks you to sign in.
{ "mcpServers": { "proxium": { "url": "https://proxium.tech/mcp" } } }
VS Code with GitHub Copilot
Add the server to .vscode/mcp.json, and sign in when VS Code asks.
{ "servers": { "proxium": { "type": "http", "url": "https://proxium.tech/mcp" } } }
Gemini CLI
Run this command. The CLI signs you in when the server asks. To sign in by hand, run /mcp auth proxium.
gemini mcp add --transport http proxium https://proxium.tech/mcp
Zed
Open Settings > AI > MCP Servers, then select Add Server and Add Remote Server. Or add the server to the settings file. Zed asks you to sign in.
{ "context_servers": { "proxium": { "url": "https://proxium.tech/mcp" } } }
Allow the connection
On the first connection, your browser opens the proxium page Connect an app. It shows the name of the client and what the client can do.
Check the name of the client before you allow it. If you did not start the connection from your client, select Deny.
- Sign in to proxium, if the page asks.
- Select Allow.
Know the tools
The server has nine tools. The project argument takes a project slug. If you are a member of one project only, the agent can leave it out.
| Tool | What it does |
|---|---|
proxium_list_projects | Lists your projects and their slugs |
proxium_recall_memory | Reads the knowledge page, the most trusted memories and the newest conversation summaries |
proxium_search_memory | Searches the memories and the past conversations by words |
proxium_read_conversation | Reads one past conversation |
proxium_query_graph | Reads the statements of the knowledge graph about one thing |
proxium_save_memory | Saves one memory: a fact, a preference, a decision, an event or a procedure |
proxium_correct_memory | Replaces the text of a memory |
proxium_rate_memory | Marks a memory useful or wrong |
proxium_forget_memory | Removes a memory |
A memory that an agent saves is an agent memory. An agent cannot correct or remove a memory that a person wrote.
The server also has the prompt proxium_load_memory. It starts a conversation with the memory of the project.
| Client | How to run the prompt |
|---|---|
| Claude Code | /proxium:proxium_load_memory |
| VS Code | /proxium.proxium_load_memory |
| Gemini CLI | /proxium_load_memory |
Install the skills
Three Agent Skills tell an agent when to read the memory and when to save to it: proxium-memory, proxium-memory-recall and proxium-memory-save.
To install them for your user, run this command. It writes them to ~/.agents/skills and ~/.claude/skills:
curl -fsSL https://proxium.tech/skills/install.sh | sh
To install them in the current repository for the whole team, run this command. It writes them to .agents/skills and .claude/skills of the current folder:
curl -fsSL https://proxium.tech/skills/install.sh | sh -s -- --project
The script replaces only the three skill folders. To read it first, open proxium.tech/skills/install.sh.
Add a project rule
A client with no skills can read a rule file. Put these lines in the file that your client reads on each request:
| Client | File |
|---|---|
| Codex, Cursor, GitHub Copilot, VS Code | AGENTS.md |
| Gemini CLI | GEMINI.md |
| Claude Code | CLAUDE.md |
## Project memory
This project has a proxium memory (MCP server "proxium").
- At the start of every task, call proxium_recall_memory and use what it returns.
- Before you assume something about the project, search with proxium_search_memory.
- When the task settles a decision, a preference, a fact or a procedure a later
session needs, save it with proxium_save_memory, one statement per call.
- Rate the memories you used with proxium_rate_memory, and correct the wrong ones.
For how memory works, refer to Memory.